Explaining 'Explainability 20%' for second competition

Alphathon
#1

What is expected for us here? For any sample should we have a value of importance for all the features, or should we have a single value of importance for every feature for the model? And how should this explaining be structured? I.e. should we make our submitted notebook take a batch of data and return predicted values and the explanations?


Thanks: Sam

#2

Hello - this question has already been answered in : https://dashboard.alphien.com/forum/262


Please ask a question on this thread if some points remains unclear. Thanks ! Lionel.